Thanks for the lecture. Is there any method to increase the precision and recall scores on the test set? I use random forest and AdaBoost based on decision tree for cross validation and it can reach over 99% of accuracy, but on the test set the result is not good. The mean accuracy is just like what you presented.
